Google announced on Friday that users can now remove the visible watermark from AI-generated images, videos, and songs. This toggle applies to the Nano Banana, Omni, and Lyria models within Gemini and the Flow video editor, with Search support arriving shortly. The update does not affect the invisible SynthID watermark or the C2PA standard metadata embedded in files. Josh Woodward, vice president for Gemini, stated this change balances creative control with safety requirements. He noted that while the visible mark is optional, the underlying transparency mechanisms remain active to help users identify synthetic content. The setting will appear under Settings and Media Watermark once the feature rolls out over the coming days. Google is also releasing a new open-source library called Credentio to help developers embed local validation tools in their applications.
This move follows Anthropic‘s decision to add watermarks to text and files generated by Claude to meet EU regulations. By separating the visual marker from the technical metadata, Google addresses practical workflow needs without removing detection capabilities. The distinction allows creators to use generated assets in professional contexts while maintaining a record of origin for verification purposes.
* Visible watermarks are optional for Nano Banana, Omni, and Lyria models
* Invisible SynthID and C2PA metadata remain unchanged
* Credentio library enables local validation for developers
